Overview
The Gridded Maps module transforms point-based data (for example, geochemical data), into continuous raster grids using machine learning. It links sample values from soil, till, or rock to auxiliary data like magnetics, gravity, or radiometrics to build a predictive model. This model estimates values across the entire Area of Interest (AOI), including areas where no samples were collected. The resulting rasters enable downstream modules like Multivariate Anomaly Maps to work with complete, gridded datasets.
This module uses predictive algorithms to find patterns between the point-based data and auxiliary data and extrapolates those patterns across space using secondary datasets.
Topic | Summary |
Module Name | Gridded Maps |
Purpose | Predicts values from point data using auxiliary rasters and ML. |
Input Format | Point-based data |
Recommended Data | Well-distributed surface datasets (soil, till, stream sediment, rock geochemistry) |
Output Format | Raster; performance scatter plot; Feature Importance graph |
Key Parameters | AOI, data column(s), modality (rock/soil/till), smoothing kernel, selected auxiliary rasters, output resolution |
Processing Summary | Trains a machine learning model on point samples and auxiliary rasters to generate stable, averaged prediction outputs across multiple model runs. |
Typical Use Cases | Creating geochem inputs for DORA, visualizing predicted distributions, supporting feature stacking workflows. |
Validation or QC | Performance scatter plot |
Common Pairings | Multivariate Anomaly Maps |
Notable Output Notes |
|
How It’s Used in Exploration
Gridded Maps help exploration teams extend the value of sparse datasets by predicting across unsampled areas. This is especially valuable in early-stage projects where coverage is incomplete, but auxiliary geophysical data is available.
Predicted rasters can be used:
As inputs for anomaly detection or machine learning targeting in DORA
To visualize potential halos or trends
To evaluate model responses in areas with no direct sampling
To build consistent input layers for feature stacking and regression workflows
Geologists should validate outputs visually and geologically, as the results are only as good as the inputs. Dense sampling and well-correlated auxiliary layers will yield stronger predictions.
Value & Benefits
The Gridded Maps module provides a practical and scalable way to convert point-based data into continuous raster layers that are compatible with machine learning and spatial modelling workflows. These predicted rasters enable downstream modules like DORA to operate effectively across the full Area of Interest, even in regions where direct sampling is limited or absent.
For example, by linking geochemical values to auxiliary datasets such as magnetics, gravity, and radiometrics, this module extends the value of existing sampling programs without requiring additional fieldwork. It supports the integration of geochemical information with other raster-based datasets, ensuring consistency across inputs used for advanced targeting, anomaly detection, and feature stacking.
The resulting maps help exploration teams visualize spatial geochemical patterns, identify trends and halos, and prioritize follow-up areas with greater confidence. Because predictions include areas of both low and high certainty, teams can use the outputs to guide fieldwork planning and risk assessment. As new data becomes available, the model can be rerun, making the module useful for iterative exploration workflows that adapt to evolving datasets and geological understanding.
Learn More
Still Have Questions?
Reach out to your dedicated DORA contact or email support@VRIFY.com for more information.